[YOLOV7X-QAT] Cannot convert onnx to trt engine

[TfeiSong]

[10/21/2024-17:11:29] [I]
[10/21/2024-17:11:29] [I] TensorRT version: 8.4.1
[10/21/2024-17:11:30] [I] [TRT] [MemUsageChange] Init CUDA: CPU +218, GPU +0, now: CPU 242, GPU 10401 (MiB)
[10/21/2024-17:11:33] [I] [TRT] [MemUsageChange] Init builder kernel library: CPU +351, GPU +327, now: CPU 612, GPU 10767 (MiB)
[10/21/2024-17:11:33] [I] Start parsing network model
[10/21/2024-17:11:33] [I] [TRT] ----------------------------------------------------------------
[10/21/2024-17:11:33] [I] [TRT] Input filename: abandon_p0.963_r0.933_v3.0_ptq.onnx
[10/21/2024-17:11:33] [I] [TRT] ONNX IR version: 0.0.7
[10/21/2024-17:11:33] [I] [TRT] Opset version: 13
[10/21/2024-17:11:33] [I] [TRT] Producer name: pytorch
[10/21/2024-17:11:33] [I] [TRT] Producer version: 1.12.0
[10/21/2024-17:11:33] [I] [TRT] Domain:
[10/21/2024-17:11:33] [I] [TRT] Model version: 0
[10/21/2024-17:11:33] [I] [TRT] Doc string:
[10/21/2024-17:11:33] [I] [TRT] ----------------------------------------------------------------
[10/21/2024-17:11:33] [W] [TRT] onnx2trt_utils.cpp:367: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[10/21/2024-17:11:33] [E] Error[3]: onnx::QuantizeLinear_2152: invalid weights type of Int8
[10/21/2024-17:11:33] [E] [TRT] ModelImporter.cpp:773: While parsing node number 9 [Identity -> "onnx::QuantizeLinear_2358"]:
[10/21/2024-17:11:33] [E] [TRT] ModelImporter.cpp:774: --- Begin node ---
[10/21/2024-17:11:33] [E] [TRT] ModelImporter.cpp:775: input: "onnx::QuantizeLinear_2152"
output: "onnx::QuantizeLinear_2358"
name: "Identity_9"
op_type: "Identity"

[10/21/2024-17:11:33] [E] [TRT] ModelImporter.cpp:776: --- End node ---
[10/21/2024-17:11:33] [E] [TRT] ModelImporter.cpp:778: ERROR: ModelImporter.cpp:180 In function parseGraph:
[6] Invalid Node - Identity_9
onnx::QuantizeLinear_2152: invalid weights type of Int8
[10/21/2024-17:11:33] [E] Failed to parse onnx file
[10/21/2024-17:11:33] [I] Finish parsing network model
[10/21/2024-17:11:33] [E] Parsing model failed
[10/21/2024-17:11:34] [E] Failed to create engine from model or file.
[10/21/2024-17:11:34] [E] Engine set up failed
&&&& FAILED TensorRT.trtexec [TensorRT v8401] # /usr/src/tensorrt/bin/trtexec --onnx=yolov7x_ptq.onnx --saveEngine=yolov7x_ptq.engine --int8

Looking forward to your reply？

[TfeiSong]

I have know the reason, and resolved it。 thanks.